Review Article Impact of Treatment on Development of NAFLD in Patients with Inflammatory Bowel Disease: a Meta-Analy

Home , Fatty liver disease, Steatohepatitis

Int J Clin Exp Med 2019;12(3):2088-2096 www.ijcem.com /ISSN:1940-5901/IJCEM0088702

Review Article Impact of treatment on development of NAFLD in patients with inflammatory bowel disease: a meta-analy

Gong Feng1, Xue-Ping Li1, Na He2, Chun-Yan Niu3, Man-Ling Liu1, Jie Gao1, Li-Ping Fan1, Ke-Lin Zhang2, Man Mi1

1Xi’an Medical University, Xi’an, Shaanxi Province, China; 2The First Affiliated Hospital of Xi’an Medical University, Xi’an, Shaanxi Province, China; 3Department of Gastroenterology, Xiangan Hospital Affiliated to Xiamen University, Xiamen, China Received November 20, 2018; Accepted January 8, 2019; Epub March 15, 2019; Published March 30, 2019

Abstract: Background: The prevalence of NAFLD has been estimated to reach as high as 40% in patients with inflammatory bowel disease (IBD), but the risk factors for the development of NAFLD in IBD remains unclear. Aim: This study aimed to assess whether IBD treatment with steroids, anti-TNF drugs, methotrexate, and the surgical treatment of IBD are associated with increased risk of NAFLD. Methods: A systematic literature review was carried out using various existing online databases from inception to August 2018, in order to identify risk factors for the development of NAFLD in IBD. The meta-analysis was conducted using the random effect model for heterogeneous data and the fixed effect model for homogeneous data. Results: The meta-analysis included seven studies, which involved 1,645 patients. Drug therapy for IBD had no significant impact on the development of NAFLD. The pooled odds ratios for NAFLD with the use of steroids, anti-TNF drugs and methotrexate were 1.31 (95% CI: 0.79-2.15, P = 0.29), 0.86 (95% CI: 0.65-1.13, P > 0.05) and 1.28 (95% CI: 0.83-1.95, P > 0.05). The odds ratio for NAFLD with IBD-related surgical treatment was 1.50 (95% CI: 1.09-2.05, P < 0.05), which was suggestive of a significant association. Conclusion: This meta-analysis revealed that steroid therapy, anti-TNF drugs and methotrexate were not associated with increased risk of NAFLD in IBD patients. In contrast, IBD-related surgical treatment was associated with elevated risk for NAFLD.

Keywords: Inflammatory bowel disease (IBD), nonalcoholic fatty liver disease (NAFLD), meta-analysis

Background However, the pathophysiological mechanisms behind the development of NAFLD in IBD pa- Inflammatory bowel disease (IBD) is a chronic tients remain poorly understood. In addition, inflammatory disease associated with multi- there are no guidelines for NAFLD detection, organ involvement. It has a tendency for recur- prevention and treatment in IBD patients. Hen- rence that requires long-term treatment, and ce, there is a need to conduct an in-depth study involves high cost The prevalence and inci- to understand the association between IBD dence of IBD in North America is the highest in and NAFLD. the world, causing huge medical burden [1]. Elevated transaminase levels are common in NAFLD is a major public health problem that IBD patients, with nonalcoholic fatty liver dis- affects approximately one billion people world- ease (NAFLD) being the most common cause wide [9]. In a society dominated by Western [2]. Recent studies have found that the preva- dietary patterns, NAFLD has been considered lence of NAFLD in IBD patients is higher than to be one of the most common liver diseases, in the general population [3, 4]. Cross-sectional resulting in a huge clinical and economic bur- studies have reported that the prevalence of den [10]. The occurrence of NAFLD in IBD pa- NAFLD in IBD can reach between 8% and 40% tients appears to be multifactorial. However, [5, 6]. Liver fibrosis has been reported in 6.4% the specific risk factors for the development of to 10% of IBD patients [7]. A recent study [8] NAFLD in IBD patients remain unclear. In the also suggested that colitis- and IBD-like ph- present metaanalysis, studies that reported enotypes are strongly associated with NAFLD. the impact of steroid therapy, anti-TNF drugs, NAFLD in inflammatory bowel disease

Figure 1. PRISMA diagram of patients with IBD, including the literature search. both UC and CD; (3) studies that used the NAFLD diagnostic criteria to define nonalcoholic fatty liver disease; (4) studies that provided a detailed description of the medical and surgical therapy for IBD.

Exclusion Criteria: (1) studies that included patients with suspected or confirmed viral hepatitis, autoimmune liver disease, or drug-induced liver disease, including alcohol related liver disease; (2) studies on IBD that failed to differentiate NAFLD and non-NAFLD patients; (3) basic medical research, review articles, or case reports; (4) methotrexate and IBD-related surgery on the studies with unclear data or different research occurrence of NAFLD in IBD patients were col- contents. lectively analyzed, in order to determine whether these are risk factors for the development of Data analysis NAFLD in IBD patients. Revman software 5.2 was used for data pro- Materials and methods cessing and analysis. The binary data variables were described as odds ratio (OR) and 95% Data sources and retrieval confidence interval (CI). The heterogeneity test was conducted using I2, where I2 > 50% was The literature review was carried out electroni- considered heterogeneous. If the data was het- cally through PubMed, Embase, the Cochrane erogeneous, the random effect model was us- Library English database, Wanfang, CNKI, and ed for the analysis, and when it was homoge- the Wipe Chinese database from the inception neous, the fixed effect model was used for the of the database to August 2018. All retrieved analysis. The funnel plot was used to identify English and Chinese articles were included publication bias. Two independent reviewers in the analysis. The following keywords were (CE and CK) assessed the risk of bias, accord- used: NASH (non-alcoholic steato-hepatitis), ing to PRISMA recommendations. Subgroup NAFLD, IBD, UC (ulcerative colitis), CD (Crohn’s and sensitivity analyses were used to analyze disease), steroids, methotrexate, tumor necro- the sources of heterogeneity. sis factor inhibitor, and histology. All human studies were included without language restri- Results ctions. Search results Inclusion and exclusion criteria A total of 416 articles were initially identified The following criteria were used to include the from the databases. Among these, 150 articles studies in the meta-analysis: (1) randomized were excluded due to duplication of data, while controlled clinical trials, cross-sectional, co- 259 articles were excluded, because these did hort, case-control, or observational studies th- not satisfy the inclusion criteria. Finally, seven at investigated the link between IBD treatment studies with a total of 1,645 patients were and NAFLD in adults; (2) studies that enrolled included in the meta-analysis (Figure 1). The

2089 Int J Clin Exp Med 2019;12(3):2088-2096 NAFLD in inflammatory bowel disease

Table 1. Basic characteristics of the included studies Study sample Study sam- Average age Proportion of NAFLD diagnos- Type of Author Year Country (NAFLD in IBD ple (IBD (year, NAFLD men (NAFLD tic means study patients) patients) in IBD in IBD patient) Sourianarayanane A [11] 2013 USA 76 217 Abdominalimaging 37.2 31% Retrospective (CT, US, MRI) Schroder T [12] 2015 Germany 30 259 Abdominal US 44.1 33% Retrospective Bessissow T [13] 2016 Canada 108 321 Histology 37.7 47% Retrospective Bosch DE [14] 2017 USA 29 49 Liver biopsy 4.7 51% Retrospective Glassner K [15] 2017 USA 56 112 Liver imaging (CT, 45 34% Retrospective US, MRI) Sagami S [16] 2017 Japan 66 303 Abdominal US 28.2 53% Retrospective Saroli Palumbo C [17] 2018 Canada 126 384 Imaging (Transient 50.2 65% Prospective Elastography) Cohort Study

analysis included six retrospective studies and one prospective study. The sub- plots of quality evaluation are presented in Figure 2A. The general plots of quality evaluation are presented in Figure 2B.

Results of the meta-analysis

Effects of steroid treatment on NAFLD: All seven studies described the impact of steroid treatment on NAFLD. The heterogeneity test revealed significant heterogeneity (I2 = 61%). The random effect model was used. The combined OR value was 1.31 (95% CI 0.79-2.15), but the difference was not statistically significant P ( = 0.29) (Figure 3A). The funnel plot revealed no publication bias (Figure 4A).

Effect of anti-TNF drugs: Six articles specifically described the effect of anti-TNF drugs on NAFLD. The hetero- Figure 2. Evaluation of the quality of included studies: (A) Sub-plots of the quality evaluation; (B) General plots of the quality evaluation. geneity test revealed no heterogeneity (I2 = 11%). The fixed effect model was used. basic features of these seven included studies The combined OR value was 0.86 (95% CI: are presented in Table 1. 0.65-1.13), but the difference was not statistically significant P( > 0.05) (Figure 3B). Funnel Quality assessment of the included articles plot revealed no publication bias (Figure 4B).

All trials were screened according to the inclu- Effect of methotrexate: Three studies specifi- sion and exclusion criteria. The present meta- cally described the effect of methotrexate on

2090 Int J Clin Exp Med 2019;12(3):2088-2096 NAFLD in inflammatory bowel disease

Figure 3. Forest plot for assessing the association between different IBD therapies and risk of NAFLD: (A) steroid treatment, (B) anti-TNF drugs, (C) methotrexate, and (D) previous IBD-related surgery.

NAFLD. The heterogeneity test revealed no het- Subgroup and sensitivity analysis erogeneity (I2 = 37%). The fixed effect model 2 was used. The combined OR value was 1.28 Heterogeneity (I = 61%) was found in the analy- (95% CI: 0.83-1.95), but the difference was not sis of the association between steroid and NAFLD in IBD patients. Six articles were elimi- statistically significant P ( > 0.05) (Figure 3C). nated one by one for the sensitivity analysis. It The funnel plot revealed no publication bias was found that when the study conducted by (Figure 4C). Sourianarayanane A et al. [11] was excluded, the heterogeneity decreased from I2 = 61% to Association between the surgical treatment for I2 = 36%, indicating that this study was the IBD and NAFLD: Three articles described the main cause of heterogeneity. Further analysis effect of IBD-related surgical treatment on revealed that this study was a case-control NAFLD. The heterogeneity test revealed no study, in which there was uniformity in the di- 2 heterogeneity (I = 0%). The fixed effect model agnostic method for NAFLD, and this study was used. The combined OR value was 1.50 included abdominal B-ultrasound, computed (95% CI: 1.09-2.05), and the difference was tomography (CT), and magnetic resonance im- statistically significant (P < 0.05) (Figure 3D). aging (MRI). However, the duration and fre- The funnel plot revealed no publication bias quency of steroid therapy and anti-TNF drugs (Figure 4D). were not detailed in this study.

2091 Int J Clin Exp Med 2019;12(3):2088-2096 NAFLD in inflammatory bowel disease

Figure 4. Funnel plot to assess for publication bias in studies that reported the association between different IBD therapies and NAFLD: (A) steroid treatment, (B) anti-TNF drugs, (C) methotrexate, and (D) previous IBD-related surgery.

Subgroup analysis of the association between steroid therapy and NAFLD in IBD patients was performed according to the country of origin of patients in these studies (U.S. and non-U.S.) and the diagnostic tools (imaging and non-imaging) used for diagnosing NAFLD (Figures 5 and 6, respective- ly). No significant heterogeneity was found in the subgroup analysis.

Discussion

Although a variety of etiolo- gies have been proposed to explain the increased prevalence of NAFLD in IBD patients, including the pres- ence of metabolic syndrome (MS), intestinal inflammation and drugs used to treat IBD, Figure 5. Subgroup analysis of the association between steroid therapy and the underlying causes and NAFLD based on the country of origin of patients (USA and non-USA). mechanisms remain largely

2092 Int J Clin Exp Med 2019;12(3):2088-2096 NAFLD in inflammatory bowel disease

caution in IBD patients with metabolic diseases, such as type-2 dia betes mellitus and hypothyroidism.

In the present meta-analysis, anti-TNF drugs were not found to be associated with NAFLD in IBD patients (OR = 0.86, P > 0.05). Serum TNF-α levels and the corre- spond ing mRNA expression in hepatocytes of NASH patients were significantly higher than those of healthy con- trols [21]. Anti-TNF-α agents have been widely used in various inflammatory diseases, and are by far the most effective drugs for the induc- Figure 6. Subgroup analysis based on the diagnostic methods used to detect NAFLD (imaging and non-imaging methods). tion and maintenance of IBD. It has been speculated that anti TNF-α agents may unknown, and require further investigation. In be protective against NASH. Infliximab has been the present meta-analysis, steroid therapy for shown to reduce steatosis and increase insulin IBD was not found to be a risk factor for NAFLD signaling in rodents on a high-fat diet [22]. In development in IBD patients (OR = 1.31, P = addition, infliximab attenuates methionine and 0.29). However, on subgroup analysis, it was choline deficiency-induced liver inflammation, found that when grouped according to country necrosis and fibrosis in rodents [23]. Adalimu- of origin (USA and non-USA), steroid treatment mab has also been shown to have a similar may be a risk factor for NAFLD in the US popu- effect [24]. Pentoxifylline is a nonselective lation (Figure 5). Glucocorticoid analogues (GC) phosphodiesterase inhibitor that reduces TNF are commonly used as induction therapy for production, and has been reported to induce IBD. Furthermore, some patients with poorly improvements in biochemical liver enzymes in controlled IBD may need a repeated or pro- NASH patients [25]. Based on the present evi- longed use of steroids. In addition, the effects dence and the present meta-analysis, anti- on carbohydrate and lipid metabolism may lead TNF-α agents are not likely to be risk factors for to metabolic syndrome (MS) and predispose NAFLD in IBD. to the development of NAFLD. In vitro studies [18] have shown that GC may induce hepato- The present meta-analysis found no associa- cyte adipogenesis and steatosis by upregulat- tion between methotrexate therapy and the ing fatty acid synthase and acetyl coenzyme A occurrence of NAFLD in IBD patients (OR: 1.28, carboxylase 1 and 2. GC and high-fat diet in P > 0.05). Methotrexate is a folate antagonist, rodent models can also exacerbate the devel- which competitively inhibits dihydrofolate redu- opment of NAFLD and liver fibrosis [19]. How- ctase, interferes with the synthesis of purine ever, there has been a lack of clinical eviden- and pyrimidine, and produces anti-inflammato- ce linking GC and NAFLD in human studies. ry effects. It can be used as an induction and Furthermore, no prospective clinical-study has maintenance monotherapy for IBD treatment, found GC to be an independent risk factor for or combined with anti-TNFα agents [26]. The NAFLD. In a study conducted by Hubel et al., serum levels of liver enzymes may be elevated compared with the control group, there was no in 15-50% of patients with methotrexate, significant difference in plasma cortisol con- although most of them are self-limiting, and the centration in NAFLD or obese patients [20]. underlying mechanism has been speculated to Although no clear guidelines havbeen estab- be correlated to oxidative stress [27]. A retro- lished, cortico steroids should be used with spective analysis revealed that approximately

2093 Int J Clin Exp Med 2019;12(3):2088-2096 NAFLD in inflammatory bowel disease

24% of IBD patients had elevated serum liver There are some shortcomings in the present enzymes after using methotrexate. However, meta-analysis. First, the dosage and duration of obvious liver fibrosis or cirrhosis is not common, drugs used in each study was inconsistent. and occur in merely 5% of patients receiving Second, the sample size of the included articles long-term low-dose methotrexate therapy. The was small. Third, other risk factors for NAFLD correlation between methotrexate and NAFLD were not consistently studied and controlled is relatively low. In a multivariate analysis, an in the included studies. Fourth, there were in- average weekly dose of 13.1 mg of methotrex- consistencies in the tools used for diagnosing ate was shown to be an independent predictor NAFLD. Future multicenter and larger sample of NAFLD [28] in patients with rheumatoid size prospective studies are required to verify arthritis, but methotrexate use has not been the results of the present meta-analysis. shown to cause NAFLD in IBD patients. Methotrexate is prone to increase liver enzymes, In conclusion, the existence of NAFLD in IBD is but does not necessarily lead to the occurrence increasingly being recognized. This is partly correlated to the increased incidence of MS and of NAFLD. complex IBD-related factors. The real impact The present meta-analysis revealed that IBD- of IBD therapy on the development of NAFLD related surgical treatment is a risk factor for needs further evaluation through prospective NAFLD development in IBD patients (OR: 1.50, studies. Furthermore, the long-term prognosis P < 0.05). One possible explanation for this of IBD patients with NAFLD requires further finding is that IBD patients undergoing surgery qualitative evaluation. In addition, there is a oftenneed parenteranutrition (PN) after surgi- lack of guidance for appropriate screening cal resection, postoperatively. Hepatic steatosis tools and strategies for NAFLD in IBD patients. is a common complication of PN [29], which Addressing these problems may facilitate early can occur in less than five days after the initia- intervention and improve prognosis. tion of PN. Hepatic steatosis may progress to Disclosure of conflict of interest NAFLD with prolonged total parenteral nutrition (TPN), coexisting MS, and the use of hepatoto ic None. drugs in IBD patients. Address correspondence to: Man Mi, Xi’ an Medi- The focus of treatments for NAFLD is generally cal University, No. 74, Hanguang North Road, Xi’an dietary and lifestyle changes aimed for achiev- 710021, Shaanxi Province, China. E-mail: a1774- ing weight loss, but these treatments have not [email protected]; Na He, The First Affiliated been specifically evaluated in the IBD popula- Hospital of Xi’an Medical University, Xi’an 710021, tion. Prevention or reversal of liver fibrosis sh- Shaanxi Province, China. E-mail: [email protected] ould ultimately lead to a reduction in NAFLD related complications. No drug preparation has References been formally approved for NASH treatment. Pioglitazone (a peroxisome proliferator-activat- [1] Rocchi A, Benchimol EI, Bernstein CN, Bitton A, Feagan B, Panaccione R, Glasgow KW, Fer- ed receptor agonist), Vitamin E and obeticholic nandes A, Ghosh S. Inflammatory bowel dis- acid (a farnesol X-receptor agonist) can improve ease: a Canadian burden of illness review. Can the histological markers of NASH [30]. Several J Gastroenterol 2012; 26: 811-817. anti-inflammatory and anti-fibrosis agents are [2] Cappello M, Randazzo C, Bravatà I, Licata A, also being actively studied. Bariatric surgery Peralta S, Craxì A, Almasio PL. Liver function may also lead to the improvement of NASH in test abnormalities in patients with inflamma- morbidly obese patients. However, it remains tory bowel diseases: a hospitabased survey. unclear whether there is a need to change the Clin Med Insights Gastroenterol 2014; 7: 25- drug therapy for IBD in patients who develop 31. NAFLD. In addition, NAFLD treatment has not [3] Principi M, Iannone A, Losurdo G, Mangia M, Shahini E, Albano F, Rizzi SF, La Fortezza RF, been specifically studied in the IBD population. Lovero R, Contaldo A, Barone M, Leandro G, Hence, at present, the treatment of these Ierardi E, Di Leo A. Nonalcoholic fatty liver dis- patients should be individualized and guided ease in inflammatory bowel disease: preva- through existing protocols for non-IBD NAFLD lence and risk factors. Inflamm Bowel Dis patients. 2018; 24: 1589-1596.

2094 Int J Clin Exp Med 2019;12(3):2088-2096 NAFLD in inflammatory bowel disease

[4] Gaidos JKJ, Fuchs M. Increased prevalence of bowel disease. Inflammatory Bowel Diseases NAFLD in IBD patients. Dig Dis Sci 2017; 62: 2017; 23: 998. 1362. [16] Sagami S, Ueno Y, Tanaka S, Fujita A, Hayashi [5] Thin LW, Lawrance IC, Spilsbury K, Thin LW, R, Oka S, Hyogo H, Chayama K. Significance of Lawrance IC, Spilsbury K, Kava J, Olynyk JK. non-alcoholic fatty liver disease in Crohn’s dis- Detection of liver injury in IBD using tran sient ease: a retrospective cohort study. Hepatol elastography. J Crohns Colitis 2014; 8: 671-7. Res 2017; 47: 872-881. [6] Bargiggia S, Maconi G, Elli M, Molteni P, Ardiz- [17] Saroli Palumbo C, Restellini S, Chao CY, Aruljo- zone S, Parente F, Todaro I, Greco S, Manzion- thy A, Lemieux C, Wild G, Afif W, Lakatos PL, na G, Bianchi Porro G. Sonographic prevalence Bitton A, Cocciolillo S, Ghali P, Bessissow T, Se- of liver steatosis and biliary tract stones in pa- bastiani G. Screening for nonalcoholic fatty liv- tients with inflammatory bowel disease: study er disease in inflammatory bowel diseases: a of 511 subjects at a single center. J Clin Gas- cohort study using transient elastography. In- troenterol 2003; 36: 417-20. flamm Bowel Dis 2019; 25: 124-133. [7] Verma S, Jensen D, Hart J, Mohanty SR. Predic- [18] Dolinsky VW, Douglas DN, Lehner R, Vance DE. tive value of ALT levels for non-alcoholic steato- Regulation of the enzymes of hepatic micro- hepatitis (NASH) and advanced fibrosis in non- somal triacylglycerol lipolysis and re-esterifica- alcoholic fatty liver disease (NAFLD). Liver Int tion by the glucocorticoid dexamethasone. Bio- 2013; 33: 1398-405. chem J 2004; 378: 967-974. [8] Chandrashekaran V Seth RK, Dattaroy D, Al- [19] D’souza AM, Beaudry JL, Szigiato AA, Trumble hasson F, Ziolenka J, Carson J, Berger FG, Kaly- SJ, Snook LA, Bonen A, Giacca A, Riddell MC. anaraman B, Diehl AM, Chatterjee S. HMGB Consumption of a high-fat diet rapidly exace- RAGE pathway drives peroxynitrite signaling- bates the development of fatty liver disease induced IBD-like inflammation in murine non- that occurs with chronically elevated glucocor- alcoholic fatty liver disease. Redox Biol 2017; ticoids. Am J Physiol Gastrointest Liver Physiol 13: 8-19. 2012; 302: 850-863. [9] LOOMBA R. Role of imaging -based biomarkers [20] Hubel JM, Schmidt SA, Mason RA, Haenle in NAFLD: recent advances in clinical applica- MM, Oeztuerk S, Koenig W, Boehm BO, Kratzer tion and future research directions. J Hepatol W, Graeter T, Flechtner-Mors M; EMIL-Study 2018; 68: 296-304. Group. Influence of plasma cortisol and other [10] Cobbina E, Akhlaghi F. Non-alcoholic fatty liver laboratory parameters on nonalcoholic fatty disease (NAFLD) pathogenesis, classification, liver disease. Horm Metab Res 2015; 47: 479- and effect on drug metabolizing enzymes and 484. transporters. Drug Metab Rev 2017; 49: 197- [21] Ruiz AG, Casafont F, Crespo J, Cayón A, Mayoga 211. M, Estebanez A, Fernadez-Escalante JC, Pons- [11] Sourianarayanane A, Garg G, Smith TH, Butt Romero F. Lipopolysaccharide-binding protein MI, McCullough AJ, Shen B. Risk factors of non- plasma levels and liver TNF-alpha gene expres- alcoholic fatty liver disease in patients with in- sion in obese patients: evidence for the po- flammatory bowel disease. J Crohns Colitis tential role of endotoxin in the pathogenesis 2013; 7: e279-285. of non-alcoholic steatohepatitis. Obes Surg [12] Schröder T, Schmidt KJ, Olsen V, Möller S, 2007; 17: 1374-1380. Mackenroth T, Sina C, Lehnert H, Fellermann [22] Barbuio R, Milanski M, Bertolo MB, Saad MJ, K, Büning J. Liver steatosis is a risk factor for Velloso LA. Infliximab reverses steatosis and hepatotoxicity in patients with inflammatory improves insulin signal transduction in liver of bowel disease under immunosuppressive tre- rats fed a high-fat diet. J Endocrinol 2007; atment. Eur J Gastroenterol Hepatol 2015; 27: 194: 539-550. 698-704. [23] Koca SS, Bahcecioglu IH, Poyrazoglu OK, Ozer- [13] Bessissow T, Le NH, Rollet K, Afif W, Bitton A, can IH, Sahin K, Ustundag B. The treatment Sebastiani G. Incidence and predictors of Non- with antibody of TNF-alpha reduces the inflam- alcoholic fattl iver disease by serum biomark- mation, necrosis and fibrosis in the non-alco- ers in patients with inflammatory bowel dis- holic steatohepatitis induced by methionine- ease. Inflamm Bowel Dis 2016; 22: 1937-1944. and choline-deficient diet. Inflammation 2008; [14] Bosch DE, Yeh MM. Primary sclerosing cholan- 31: 91-98. gitis is protective against nonalcoholic fatty [24] Yalcin M, Akarsu M, Celik A, Sagol O, Tunali S, liver disease in inflammatory bowel disease. Ertener O, Bengi G, Akpinar H. A comparison of Hum Pathol 2017; 69: 55-62. the effects of infliximab, adalimumab, and [15] Glassner K, Malaty HM, Abraham B. Epidemiol- pentoxifylline on rats with non-alcoholic stea- ogy and risk factors of nonalcoholic fatty liver tohepatitis. Turk J Gastroenterol 2014; 1: 167- disease among patients with inflammatory 175.

2095 Int J Clin Exp Med 2019;12(3):2088-2096 NAFLD in inflammatory bowel disease

[25] Adams LA, Zein CO, Angulo P, Lindor KD. A pilot [29] Angelico M, Della Guardia P. Review article: trial of pentoxifylline in nonalcoholic steato- hepatobiliary complications associated with hepatitis. Am J Gastroenterol 2004; 99: 2365- total parenteral nutrition. Aliment Pharmcol 2368. Ther 2000; 14: 54-57. [26] Dignass A, Van Assche G, Lindsay J, Lémann [30] Natta ML, Abdelmalek MF, Chalasani N, M, Söderholm J, Colombel JF, Danese S, Dasarathy S, Diehl AM, Hameed B, Kowdley D’Hoore A, Gassull M, Gomollón F, Hommes KV, McCullough A, Terrault N, Clark JM, Tonas- DW, Michetti P, Morain C, Oresland T, Windsor cia J, Brunt EM, Kleiner DE, Doo E; NASH Clini- A, Stange EF, Travis SP; European Crohn’s Coli- cal Research Network. Farnesoid X nuclear re- tis Organisation (ECCO). The second European ceptor ligand obeticholic acid for non-cirrhotic, evidence-based consensus on the diagnosis non-alcoholic steatohepatitis (FLINT): a multi- and management of Crohn’s disease: current centre, randomised, placebo-controlled trial. management. J Crohns Colitis 2010; 4: 28-62. Lancet 2015; 385: 956-965. [27] Bath RK, Brar NK, Forouhar FA, Wu GY. A rview of methotrexate-associated hepatotoxicity. J Dig Dis 2014; 15: 517-524. [28] Sakthiswary R, Chan GY, Koh ET, Leong KP, Thong BY. Methotre xate-associated nonalcoholic fatty liver disease with transaminitis in rheumatoid arthritis. Scientific World Journal 2014; 2014: 823763.

2096 Int J Clin Exp Med 2019;12(3):2088-2096